List:General Discussion« Previous MessageNext Message »
From:Michael Widenius Date:June 8 1999 6:22pm
Subject:SQL_SMALL_RESULT
View as plain text  
>>>>> "Mark" == Mark Papadakis <markp@stripped> writes:

Mark> Hi!
Mark> Let's say we create a ramdisk.
Mark> And we use --tmpdir /ramdisk when starting the mysql daemon.
Mark> Will that be faster in group by operations or shall I stick to using
Mark> SQL_SMALL_RESULT?

HI!

SQL_SMALL_RESULT should be much faster;  MySQL uses hashing in this
case, which is much faster than b++ trees.

Regards,
Monty
Thread
SQL_SMALL_RESULTMark Papadakis8 Jun
  • SQL_SMALL_RESULTMichael Widenius8 Jun